Type I error rates were estimated for three tests that compare means by using data from two independent samples: the independent samples t test, Welch's approximate degrees of freedom test, and James's second order test. Type I error rates were estimated for skewed distributions, equal and unequal variances, equal and unequal sample sizes, and a…

Describes and compares eight methods of estimating the parameters of the Zipf distribution. (CLB)

Two methods of estimating parameters in the Rasch model are compared. The equivalence of likelihood estimations from the model of G. J. Mellenbergh and P. Vijn (1981) and from usual unconditional maximum likelihood (UML) estimation is demonstrated. Mellenbergh and Vijn's model is a convenient method of calculating UML estimates. (SLD)

This paper considers the following questions: (1) what is the relationship between the method of paired comparisons and Rasch measurement theory? (2) what is the relationship between the method of paired comparisons and graph theory? and (3) what can graph theory contribute to the understanding of Rasch measurement theory? It is specifically shown…

Log-linear model bivariate smoothing and a bivariate smoothing model based on the four-parameter beta binomial model were compared for usefulness in frequency estimation common-item equipercentile equating using two datasets. The performance of smoothed equipercentile methods was also compared to that of linear methods of common-item equating.…

Using computer simulated data, the Type I error rate and statistical power were empirically estimated for several pairwise multiple comparison strategies for situations where population variances differ. Focus was on comparing modified Bonferroni procedures with Dunnett's solutions, and determining whether or not J. P. Shaffer's suggestion of…

The purpose of this paper is to illustrate the use of tobit analysis and tobit decomposition in educational research. Tobit estimates of growth rate in the presence of a ceiling effect were compared with ordinary least squares (OLS) and weighted least squares (WLS) estimates. The tobit estimates had the smallest standard error, the smallest bias,…

In experiments with paired comparisons, judges are occasionally allowed to express indifference between alternatives. For the analysis of such data, models for paired comparisons with ties are needed. Several models with tie parameters are reviewed. All are extensions of the basic models of L. L. Thurstone (1927) and R. A. Bradley and M. E. Terry…

A method of comparing one-step M-estimates of location for heavy tailed distributions is proposed and investigated. Simulations indicate that the new procedure provides good control over Type I errors and has more power than do some other methods for dealing with heavy tailed distributions. (SLD)

E.J. Bedrick recently derived the asymptotic distribution of the modified sample biserial correlation estimator of F.M. Lord and studied its efficacy for bivariate normal populations. A more detailed examination of the properties of Lord's estimator is provided, and the small sample bias and variance are studied. (SLD)

Across conditions, differences between population parameters defined by common factor analysis and component analysis are demonstrated. Implications for data analytic and theoretical issues related to choice of analytic model are discussed. Results suggest that principal components analysis should not be used to obtain parameters reflecting latent…

Contingency tables, and their associated statistical tests, are frequently used in educational and social research. Popular statistical tests used in contingency table analyses include the Pearson chi-square test and the likelihood ratio chi-square test. These two tests are chi-square distributed under large sample conditions. However, when a…

The influence of sample size (N) and model parsimony on a set of 22 goodness of fit indices was investigated, including those typically used in confirmatory factor analysis and some recently developed indices. For sample data simulated from 2 known population data structures, values for 6 of 22 fit indices were reasonably independent of N and were…

An investigation of the effects of randomly missing data in two-predictor regression analyses is described. The differences in the effectiveness of five common treatments of missing data on estimates of R-squared values and each of the two standardized regression weights is also investigated. Bootstrap sample sizes of 50, 100, and 200 were drawn…

During the past several decades, a new class of measurement theory, called latent trait (or item response) theory, has become popular. Supporters contend that results obtained through the use of this theory prove vastly superior to classical measurement statistics. This paper investigates the use of the Rasch model, one of the more popular latent…